@fro.bot/systematic 3.26.4 → 3.27.1

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
@@ -17,9 +17,7 @@ For description-only updates, follow the Description Update workflow below. Othe
17
17
 
18
18
  ## Context
19
19
 
20
- **If you are not OpenCode**, skip to the "Context fallback" section below and run the command there to gather context.
21
-
22
- **If you are OpenCode**, the six labeled sections below contain pre-populated data. Use them directly -- do not re-run these commands.
20
+ The six labeled sections below (Git status, Working tree diff, Current branch, Recent commits, Remote default branch, Existing PR check) may be filled in by the host before you read them. If any section is empty, or still shows its raw command line instead of output, the host did not run the commands: run the command in the "Context fallback" section below and use its output instead. Otherwise, use the data directly throughout this skill and do not re-run these commands.
23
21
 
24
22
  **Git status:**
25
23
  !`git status`
@@ -41,7 +39,7 @@ For description-only updates, follow the Description Update workflow below. Othe
41
39
 
42
40
  ### Context fallback
43
41
 
44
- **If you are OpenCode, skip this section — the data above is already available.**
42
+ **Run this only when the sections above are empty or still show raw command lines.**
45
43
 
46
44
  Run this single command to gather all context:
47
45
 
@@ -110,7 +108,7 @@ Report the PR URL.
110
108
 
111
109
  ### Step 1: Gather context
112
110
 
113
- Use the context above. All data needed for this step and Step 3 is already available -- do not re-run those commands.
111
+ Use the context above, or the output of the Context fallback command if the sections were empty. This covers this step and Step 3. Do not re-run commands whose output you already have.
114
112
 
115
113
  The remote default branch value returns something like `origin/main`. Strip the `origin/` prefix. If it returned `DEFAULT_BRANCH_UNRESOLVED` or a bare `HEAD`, try:
116
114
 
@@ -0,0 +1,75 @@
1
+ # OpenCode v2 Capability Profile
2
+
3
+ Evidence registry: see [`HARNESSES.md`](../../../HARNESSES.md).
4
+
5
+ | Capability | Mechanism | Status | Fallback |
6
+ |---|---|---|---|
7
+ | Subagent delegation | `subagent` tool; `agent` selects a specialist, `sessionID` resumes a session, `background: true` runs it asynchronously | supported; foreground by default | Dispatch in the foreground, serially or in small batches |
8
+ | Blocking user interaction | `question` tool | supported | Present numbered options in chat and wait |
9
+ | Task tracking | `todowrite` tool | conditional: only offered when the host's experimental `todo` option is on | Maintain a visible task list in responses |
10
+ | Skill loading | `systematic_skill` for bundled skills, with arguments; `skill` tool (by `id`) for non-Systematic skills | supported | Read the skill instructions listed by the active harness |
11
+
12
+ ## Invocation examples
13
+
14
+ ### Subagent delegation
15
+
16
+ ```typescript
17
+ subagent({
18
+ agent: "systematic-implementer",
19
+ description: "Implement auth module",
20
+ prompt: "Implement the auth module and return the changed files.",
21
+ })
22
+
23
+ // Background dispatch (a background run returns immediately; its result is available later):
24
+ subagent({
25
+ agent: "repo-research-analyst",
26
+ description: "Map fixture patterns",
27
+ prompt: "…",
28
+ background: true,
29
+ })
30
+
31
+ // Resume a prior specialist session:
32
+ subagent({ agent: "systematic-implementer", sessionID: "<prior-session-id>", description: "Resume the implementation task", prompt: "…" })
33
+ ```
34
+
35
+ Skill text that names OpenCode's `task` tool with `subagent_type` means the `subagent` tool with `agent` on this host, and `task_id` means `sessionID`.
36
+
37
+ `agent` takes a bare agent stem — OpenCode registers bundled agents under
38
+ their filename stem, not a namespaced identifier. Prose that *refers* to an agent
39
+ uses the canonical `systematic:<category>:<name>` form instead, which the
40
+ content-integrity gate validates against real agent files. Leave `model` unset
41
+ unless the user names a model.
42
+
43
+ Both forms are enforced. Examples in bundled content must name agents this
44
+ package actually ships, so they work for every reader. A host may provide
45
+ additional agents through the user's own configuration; those are available at
46
+ runtime but must not appear in bundled examples.
47
+
48
+ ### Blocking user interaction
49
+
50
+ ```typescript
51
+ question({
52
+ questions: [{ question: "Which deployment target should I use?", header: "Deployment target",
53
+ options: [{ label: "Staging", description: "Deploy to staging" }] }],
54
+ })
55
+ ```
56
+
57
+ ### Task tracking
58
+
59
+ ```typescript
60
+ todowrite({
61
+ todos: [{ content: "Run validation", status: "pending" }],
62
+ })
63
+ ```
64
+
65
+ `status` is one of `pending`, `in_progress`, `completed`, `cancelled`. When `todowrite` is not offered, keep the task list in your responses instead.
66
+
67
+ ### Skill loading
68
+
69
+ ```typescript
70
+ systematic_skill({ name: "systematic:using-systematic" })
71
+ systematic_skill({ name: "systematic:reproduce-bug", arguments: "ISSUE-123" })
72
+ skill({ id: "external-skill" })
73
+ ```
74
+
75
+ Pass `arguments` to `systematic_skill` only when the user supplied them; the `skill` tool takes no arguments.